EMAX 250 Pro FPV mini build (with Naze32 GPS experiments)

[jhitesma]

And short version of what it took to get to that point

Detailed post to follow late tonight. Going to go splash in the pool for a bit with my girls then get some dinner and get the little one off to bed, then I can dig back in on the next few steps. So tempting to just go fly it now though

 

[midnightdaylight]

(Never mind... lol I just saw it above my post...)


Do you have a youtube channel by any change? I would like to sub for videos of this thing =D

 

[jhitesma]

Given that my videos are more or less unedited I'm always shocked when someone subscribes to my channel

One of these days I'll get my good computer back here so I can do some actual editing. My wife is the editor I'm just a camera man

Before we ate dinner I strapped the mobius on the mount to give it a try. Not bad. Lot of jello sitting on the ground but once I got airborne it smoothed out and I think it was mostly just some slight oscillation in the quad I was seeing. I may try it without the mount, I was surprised how well no mount worked on the ultrabudget, the twitchity and the warp and this feels smoother than the two DYS setups but not as smooth as the warp setup so I'd expect decent results from it even without the dampening balls.

Also I looked at the instructions that came with the ESC's and it looks like that whine is the default settings. Instructions say the control frequency is 8Khz but switchable to 22KHz "This option is the drive frequency of the motors" wonder if i can program all of them at once using the motor slider in baseflight. Might look into what it takes to do an initial flash to blheli

I also see timing is defaulted to medium and a low voltage cutout enabled. I may try hooking them up to see if BLHeli suite can find a bootloader. Reprogramming all of that through sticks 4 times if I can't do it through the motor slider just sounds like torture. I'd rather just reflash them to bl and be done with it. So trying through BL first seems like a decent idea

 

[jhitesma]

In regards to my remarks in the video about the lack of grass. It's totally true, and even if my lawn was grown in I'd need some kind of platform because I grow my grass too tall to take these mini's off from! But right now my entire backyard has been taken over by our butternut squash plant. Just did our second harvest Saturday evening and there's still about this many again on there and just as many new flowers still. Which is a shame because by the end of the month I plan on tilling it under and reseeding for our winter lawn. I'd like to let it keep producing but it's just taking way to much of our yard:



Ok - now on to the build

Once I had the motors back on the arms I went ahead and mounted the arms up. Of course after I did this I realized I'd have to run the wires on the bottoms of the arms instead of the tops so I had to take take the motors back off to get the wires through the close holes without stressing them too much. That done I finally had the base and arms assembled and decided to tackle mounting the boards. I decided trying to cut standoffs was bound to fail, and didn't have any luck finding the sizes I wanted at my local stores. So...I looked for something else to use.

Then it hit me. Prop adapters! I don't keep all of my unused prop adapters but I keep a handfull or two because they do sometimes come in handy. I grabbed four 3mm adapters and they were perfect:


The prop adapters give it just enough room for wires to clear on either side without much danger of shorting out. Once it's fully wired I'll give this board a coating of liquid electrical tape just to be safe. But it looks like plenty of clearance unless I really crash hard:


Side view shows the relative mount sizes, I'd still like to find shorter posts for the Naze, I really don't need that space under it and more room above it will help me put my FPV gear where I want to put it:


I can work with it though. There's enough room for everything - just may not be able to mount it all right where my first choice would be. Might fit...we'll find out soon enough:


Time to mount the ESC's. This is when I found out my time lapse camera was shutting down. Which thankfully reminded me to get some photos because I completely forgot to get shots of prepping the ESC's. On three of them I gently and carefully cut open the heat shrink just enough to desolder the power wires from the servo cable. Then I added a drop of hot glue to help protect the lone wire where it connects and hold the heat shrink closed again.

The first one didn't come out very great, but by the 3rd one you could barely tell it was modified other than the two missing wires. I tried to be careful and leave the wires as three as long as possible so I can re-use them on other things in the future.


I then test fit the ESC's and proceeded to trim their power wires down and mount them up. I mounted them all with the FET's facing "out" or down which is actually up in the photos since I'm working on it upside down. To mount them I first put a small drop of hot glue on the inner edge of a cutout where I was going to pass any wires from the ESC - let it cool about halfway and then gently press the wire in. I then used a little bit of hot glue to hold the ESC's against the frame. This may melt when they get warm...but it's mostly to keep them in place while building. Once it's built the wires are all short enough and the bottom captures them well enough they aren't going anywhere even if the glue melts:


The ESC's don't quite fit side by side, but there's enough vertical clearance they can overlap a tiny bit without causing any issues. Two down and looking good! I really took my time routing the signal wires (and the two power wires I left intact) and protecting them with bits of hot glue even though I had sanded the holes. Just want to make sure they don't rub and cause problems:

 

[x0054]

You can cut the posts for the Naze, and then take a M3 screw, cut a slit perpendicular to the threads and run it back and forth through the whole in the post. It will cut the necessary threads into plastic quite easily. It's a little trick that worked for me on occasions when I can not find a right size tap. Just an idea.

Other then that, looks like a really clean build.

- Bogdan

 

[jhitesma]

I really wanted to keep the wiring clean on this, so I spent a lot of time and attention to the hot glue holding the wires in place. That also meant waiting for it to cool and harden so I probably took more photos of this than really necessary. It's not like my hot glue work will win any awards - it's ugly:


I did use some of the cooling time to solder the GPS antenna to the board because I was sick of it bouncing around and didn't want it to damage the connector between the antenna and board:


I debated for a while on just how to mount the Naze. I decided I didn't want to do it in it's standard position since that puts the baro sensor towards an edge. Rotating it 90 degrees centers the baro and keeps it more shielded. This was rotated to put the USB on the right side...but I ended up changing my mind. I figured the moto board had the USB on the left and I'll probably get to help my friend tune his warp more when he comes home for break...so keeping things consistent to the left made more sense:


Another closer look at how I protected the wires. After this hardens the wires are shielded from the carbon - but can still pop loose. So I go over it with another small dab of glue to lock them in:


All four ESC's mounted and motors soldered. Note on soldering the motors. I paid careful attention to how I was doing it to try and get them right first try. I ended up close. The two counter clockwise motors I had swapped two of the wires instead of hooking them up "straight" - well they were both backwards and needed to be "straight". So the way to do this easy first time is hook the longest ESC wire (the one that reaches furthest up into the ESC) to the top wire from the motor, the middle motor wire to the middle ESC wire and the shortest ESC wire to the lowest motor wire on both sides. The rotate the quad and do the other two. Still I only had to redo 4 wires so it only took me a minute or two. And not having the bullets let me route the wires very cleanly. Now that they're all right I'm actually going to undo them and re-do them so I can add some heat shrink or sheathing (I just hate working with sheathing) to help protect them where they cross through the arm, hide the white shrink and just generally clean the bundle of 3 wires up to look like one thick wire:


Next I shortened the signal and BEC wires and soldered them on:


Got a little mixed up for a minute and forgot which way the motors were numbered, but caught myself and fixed it easily. I left enough wire I can flip the naze over to work on it, not quite as clean as I'd like due to the extra wire but I think the utility is worth it:


Told you my hot glue work isn't the cleanest. But should protect the wires well enough.


About now I started to get excited...the hardest part is done!

 

[jhitesma]

You can cut the posts for the Naze, and then take a M3 screw, cut a slit perpendicular to the threads and run it back and forth through the whole in the post. It will cut the necessary threads into plastic quite easily. It's a little trick that worked for me on occasions when I can not find a right size tap. Just an idea.

Yeah, I've done that before...but I'm really not very good at cutting threads. It's a job I usually get a friend to do if I can since I usually muck it up. And since I don't have any spare posts I'm a little nervous. I may tackle it eventually....but for now I'll live with the extra height.

Other then that, looks like a really clean build.

Thanks, Still a bit of cleanup to go and it will get worse when I add the FPV gear...but I'm working hard to keep this clean. I'm really trying to make this a nice reliable build. Funny, I've got the least funds invested in this build than any of my others - but it's getting the most care and attention because it still obviously has the highest value despite me not having to pay for it

Working on it keeps inspiring me to make the extra effort and do it right instead of just tossing it together and getting it in the air as quickly as I can like I usually tend to do.

The high winds this weekend probably didn't hurt either, the poor FT-22 didn't get to maiden at all yet

 

[jhitesma]

Ok, hard part is done...but still need a way to control this thing! I already decided to give the openLRS a go on this build so I looked for where I'd like to make it fit. I decided on the back "patio" made the most sense but experience has taught me it needs some protection. So I cut a pad of 1/4" black craft foam just slightly larger than the unit, then pressed it into the foam to show where I would have to carve out clearance for the crystal on the RF Module. I had to expand it a bit more than I expected but eventually got a shape that fit well:


I mounted it with a bit of velcro that was handy. Though I later realized that the velcro covered up the LED's on the RX making debugging harder and it tended to pull the other end of the unit up. Plus I'm nervous about velcro in general, I've broken quite a few velcro straps on the wood quad and twitchity already:


The RX connects with just 3 wires, power, ground and PPM. The first three on the naze. Super simple to wire up. Don't forget to do a "feature PPM" and then "feature" to make sure only PPM is active in the Naze CLI to activate PPM input. I added a drop of hot glue to protect the wires here as well:


It should in theory be ready to fly now! All the minimal requirements are hooked up...well, except a battery cable and battery strap - not to mention battery itself:


But I don't want to try and fly with it like that - the naze is far too exposed. And given how the maiden(s) ended up going it's a good thing I opted to install the upper deck for protection first:


I also installed the mobius mount just so I wouldn't loose it. I thought about leaving it off so it wouldn't get damaged in an accident, but decided there was bigger risk of me loosing it than me damaging it - that may or may not have been a good call:


Since it's about ready for software tweaks and test flying I tried to grab a few glamour shots since it will probably never look quite this clean again!


Out of curiosity I got a weight of it ready to fly minus battery, battery strap, and battery cable:


One note: I do plan on replacing the stock orange openLRS antenna which is really really horrible. I'm going to make a Vertical dipole that will hang off the back similar to what the VAS/StoneBlue guys have been running on their 250 H quad builds lately. Just need to talk to my friend the bike geek about getting some bicycle brake cable to make it from

Next I hooked it to the computer and did some initial setup on the Naze. Specifically I did the "feature PPM" I knew I needed, I also enabled "feature failsafe" and changed the failsafe variables to:

failsafe_delay = 10
failsafe_off_delay = 10
failsafe_throttle = 1200
failsafe_detect_threshold = 985

I have my openLRS set to drop PPM completely when it looses signal which drops the signal well below 985

I didn't do any other changes at this point. I thought I took a look at the 3D copter orientation deal in baseflight config and confirmed it was reacting correctly...which is why I totally forgot about doing "align_board_yaw = 90" at this point.

Next I went to confirm that I had the motors wired right. I may have implied in my last post that this was done already but honestly I couldn't do it yet since I didn't have a power cable yet! So that was next.

 

[jhitesma]

I decided to go ahead and use my heavier 16ga wire instead. Anything smaller I just worry about the weight of the XT60 being too much for the tiny wires. So I soldered some up and added a bit of heat shrink. Getting these soldered proved a bit tricky. The 16ga wire is pushing the limit on the holes in the afro board. I didn't overtin them so they'd fit through - but as soon as I touched the wire to the pad it acted like desoldering braid and sucked most of the solder off the pad up into the wire making it way harder to get it through the hole. So this was not as clean of a solder joint as I would have liked and is my biggest disappointment of the build so far. I had to reheat and reshape the joints twice to get them where I felt secure they wouldn't accidentally short out against the bottom plate once it was installed. Again, I plan on adding liquid electrical tape when it's done but I still have to solder on the FPV power connection and maybe some lights so I'm not ready to coat it yet:


Now, all we need is props!


With props we come to 433g ready to fly.


Looks ready to test!



This is now where the videos should actually come in, but since I shared them earlier I won't bother moving them here. My first test flight failed. As soon as I gave it a bit of throttle it just started doing flips and and fell off the wall I like to launch off of into the neighbors yard. Horray, one broken prop and two motors in the dirt before I even got to fly
:black_eyed:

As I was replacing the broken prop I realized the two black props were on the wrong side. Doh. Easy fix. Too easy. Because I spotted that so quickly I didn't bother to look for any deeper issues. I did fire it up with no props and tried to feel how it was reacting - but it was hard to judge without props. So I went for the 2nd try.

2nd try I looked for grass to launch off of..but it's all too tall to launch out of without a piece of wood or cardboard to keep the grass out of the props. And really the ground under the grass is still just sand so a flip on grass still means motors in the sand. Driveway seemed safer by comparision. And again, as soon as it started to get off the ground it went crazy. Tested it in my hand with props and then it hit me that I had never reconfigured the board for being rotated. Doh!

Hooked it back up and got it responding right on screen. Then gave it a go...and it flew! But the wind had picked up and I didn't want to fly in that kind of wind in such a tight space. And I didn't want to crash before I got the other gear on - so I only flew enough to confirm acro and horizon work and then brought it in.

I later tossed the mobius on it, but again flew very gently and only for a minute or so just to see how the footage looks.

Really happy that I met my goal of getting it flying today. I could have pushed and got the GPS and FPV gear on as wel, but I'm enjoying taking my time and doing this one right. Will go try and hookup telemetry right now but won't be tackling GPS or FPV gear tonight at this rate.

 

[Twitchity]

The quad is looking nice, and the first maiden attempt was entertaining to watch as well

If you ever figure out how to get rid of that horrid whine from the ESCs, please post a tutorial on how to do so. The sound isn't as bad in person as the videos make it out to be (at least in my case), but it's still extremely annoying.

 

[cranialrectosis]

Nice clean build, jhitesma.

I have had exactly the same issue when forgetting to set align_board_yaw = 90. Looks like a Tazmanian devil on the launch pad. :black_eyed:

Did you test failsafe? Will 1200 shut the motors down or spin them slowly?

 

[jhitesma]

The quad is looking nice, and the first maiden attempt was entertaining to watch as well

If you ever figure out how to get rid of that horrid whine from the ESCs, please post a tutorial on how to do so. The sound isn't as bad in person as the videos make it out to be (at least in my case), but it's still extremely annoying.

Changing that setting from 8khz to 22khz should do it...I tried doing it them all at once using the motor slider in baseflight...but that was a no go. So It looks like it will take doing them each individually...and if I'm going to do that kind of effort I will probably just flash them to blheli so I can make changes over the signal wire instead in the future.

For now I'll live with the whine as I've now got bigger issues...more about that in a moment.

 

[jhitesma]

Nice clean build, jhitesma.

I have had exactly the same issue when forgetting to set align_board_yaw = 90. Looks like a Tazmanian devil on the launch pad. :black_eyed:

Yeah, I told myself at least a dozen times during the build "Don't forget to set the board alignment" but even after that first flip off the wall I still forgot it :black_eyed: thankfully no major damage was done and it was a quick fix!

Did you test failsafe? Will 1200 shut the motors down or spin them slowly?

Tested it with no props. 1200 won't shut them down but it's not even high enough to maintain a hover. 1170 is where the motors start to spin right now. Plus it only stays at that for 1 second with the setting I'm using, so it's 1 second for failsafe to kick in, then 1 second of the props still spinning but at 1200 and then shut down the motors and fall from the sky if it's not already on the ground.

 

[Twitchity]

Outside of ESC calibration using the sliders in Baseflight, how do you change the khz rating? I'm still learning BF so I'm wondering if I overlooked something. I use the sliders to originally calibrate the ESC's on my acro quad but I still get that whine.

 

[jhitesma]

So last night I tried to hook up telemetry. Should have been nice and simple, just a matter of plugging my BT module into my TX module and then hooking two wires from the central RX/TX pins on the Naze32 over to the TX/RX pins on the openLRS RX. Easy Peasy. Well, no.

It didn't work. My BT module has been bashed around a bit and sometimes looses connection so that was my first suspect. Sure enough it's power light wasn't coming on. Wiggled and fiddled and figured out I had forgotten the pin order on the TX and misread the labels - had the power reversed. Doh. Swapped that around and the BT powered up and connected to my phone...but no data.

Swapped the RX/TX lines around in case I had them backwards...nope. Swapped RX/TX on the RX incase I had them backwards there...nope. Just no data. Hmmm....

Let's take the RF out of the picture. Just hooked the BT straight to the wire coming off the RX/TX on the Naze. Nope, still no data. Ok, this is odd...don't seem to be getting a MSP stream off there like I should be. Did some research and couldn't find anything about needing to enable it in CLI or anything - everyone seems to say just hook up the BT module and it just works like it does on MW.

So I decided to put that aside and instead work on calibrating the ESC's better. But wasn't having much luck with that using either the CLI method or the motor slider method. ESC's just don't seem to take a cal. Fired up my TX to arm it and see if the motor will still spin at all after the ESC tweaking I had tried...and nothing. Then I noticed my RX was just blinking and never indicating a connection. Odd.

Tried rebinding...but no go. Hmm. Not good. Decided to just go ahead and reflash the openLRS modules with the latest code. TX flashed no big deal. Then hooked up the RX and tried to flash it...no go. Serial problems. Hmm...if I had fried the RF module somehow (fairly easy to do on these openLRS modules) I should still be able to reflash the microcontroller since the RF module is the bit that's not 5v tolerant - the uC is fine at 5v. But now it seems the uC has gone flaky.

Really not sure what's going on with the RX now but it all seems to have started when I hooked up those telemetry wires. I've heard rumors about blowing RX modules with 5v serial signals - but I've always just run it straight from my MW like that with no issues...and Naze is probably running 3.3v serial anyway (I'll have to get out my scope and check tonight.) Besides, if I had blown out the RF module the uC should still respond and allow reprogramming. Something has gone horribly wrong here.

Looks like I'm going to have to order a new openLRS rx I've been looking for an excuse to for awhile so it's not a huge deal...but still stings and slows me down. I'll have to use a lemon RX with UART for now which means giving up a serial port and not being able to run telemetry - as well as not being harder to run OSD and GPS at the same time.

May just try and make a homebrew openLRS rx tonight instead.

Really puzzled over what's going on though. Then again I've been suspecting this RX of goofiness for some time which is why it's been in my parts box instead of on my knuckle. Maybe it just finally gave out.

Will update later today when I can investigate further my new mysterious RX failure.

 

[jhitesma]

Outside of ESC calibration using the sliders in Baseflight, how do you change the khz rating? I'm still learning BF so I'm wondering if I overlooked something. I use the sliders to originally calibrate the ESC's on my acro quad but I still get that whine.

Basically you should be able to use the sliders like a TX stick to do it as explained in the instructions that come with the ESC's. The problem is it seems sometimes baseflight doesn't really return to zero on those sliders as I slide them down and it shows 0 but the motors stay spinning.

The general idea is you power up with the TX already at full throttle as if you were going to do a calibration...but then you hold it there instead of dropping it to zero. Then the ESC will go into programming mode and start beeping out which setting it's on. When it beeps out the right number of beeps for the setting you want to change you move the throttle to zero to switch it.

Doing it on one ESC is a pain, doing it on four is something I'm really not looking forward to. Especially since the baseflight sliders don't seem reliable enough to do it so it may have to be done by connecting them straight to a RX to program.

If you do it I'd also suggest disabling the LVC and maybe bumping the motor timing to med-high while you're at it.

 

[Twitchity]

Basically you should be able to use the sliders like a TX stick to do it as explained in the instructions that come with the ESC's.

Being of the male gender, I immediately threw out the instructions that came with my ESCs and cut off the heatshrink to void the warranty Time to start looking online for the little piece of paper that came with my ESCs. Thank you though, now I know what I need to try.

 

[jhitesma]

I can scan mine in at lunch for you.

 

[Twitchity]

I believe I found it, but a second check wouldn't hurt. http://www.xheli.com/66p-107-simon-series-12a.html

Looks like I'll change D2 to mid-high, and D5 to 22khz, right? I'll see if BF gives me better results when I get home from work.

 

[x0054]

On the Naze32 the telemetry port is HW inverted serial that is shared with the USB port. So, in order to get telemetry from it you need to be able to receive inverted serial, the USB has to be unplugged, and the board must be armed. I found it easier to connect telemetry over soft serial, probably safer too. However, even over soft serial I can only get telemetry when the board is armed.